We evaluated the usefulness of commonly ordered routine admission laboratory tests in 301 patients admitted consecutively to the internal medicine wards of a university teaching hospital. Using a consensus analysis approach, three Department of Medicine faculty members reviewed the charts of admitted patients to determine the impact of the test results on patient care. The evaluated tests were the urinalysis, hematocrit, white blood cell count, platelet count, six-factor automated multiple analysis (serum sodium, potassium, chloride, bicarbonate, glucose, and blood urea nitrogen), prothrombin time, partial thromboplastin time, chest x-ray, and electrocardiogram. Forty-five percent of the 3,684 tests were ordered for patients without recognizable medical indications. Twelve percent of these routine tests were abnormal, 5% led to additional laboratory testing, but only 0.5% led to change in the treatment of patients. We conclude that the impact of routine admission laboratory testing on patient care is very small and that there is little justification for ordering tests solely because of hospital admission.